Hi
Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> writes:
> I like this idea substantially better. More or less as attached?
Yes, thanks for writing it.
It passed all the local tests I had before and the performance is
back.
One small thing. Looks like HJ_FILL_OUTER_NULL_TUPLES can
now only be reached in a parallel join, so maybe add
Assert(parallel) there.
Thanks,
Shihao